Linear genetic programming based on an age-layered population model

被引：0

作者：

Cao B. ^{[1
]}

Jiang Z. ^{[1
]}

Zhang J. ^{[1
]}

机构：

[1] College of Information, Beijing University of Technology, Beijing

来源：

Harbin Gongcheng Daxue Xuebao/Journal of Harbin Engineering University | 2017年 / 38卷 / 04期

关键词：

Age-layered population model; Bloat; Diversity; Genetic programming; Linear genetic programming; Over-fitting; Premature convergence; Two-layer tournament;

D O I：

10.11990/jheu.201602025

中图分类号：

TP31 [计算机软件];

学科分类号：

081202 ; 0835 ;

摘要：

To alleviate premature convergence and bloat in general linear genetic programming, a modified linear genetic programming method based on an age-layered population model is proposed. To alleviate premature optimization of the population, we first applied an age-layered population model to linear genetic programming to improve the integral population diversity. We then used a two-layer tournament to improve the sub-population diversity in each layer, improving the local population diversity and decreasing the occurrence rate of premature optimization by increasing the diversity of the population. To control the bloat effect of the population, the age-layered population model segregated individuals into different layers based on age, so the quantity of long-length individuals was limited. The experimental results on five symbolic regression benchmark functions show that the proposed method can improve population diversity to reduce premature convergence and effectively control bloat. © 2017, Editorial Department of Journal of HEU. All right reserved.

引用

页码：610 / 616

页数：6

共 19 条

[1]

Brameier M., Banzhaf W., Linear Genetic Programming, pp. 1-8, (2007)

[2]

Koza J.R., Genetic Programming: On the Programming of Computers by Means of Natural Selection, pp. 17-63, (1992)

[3]

Gandomi A.H., Danial M.S., Alavi A.H., Et al., Linear genetic programming for shear strength prediction of reinforced concrete beams without stirrups, Applied Soft Computing, 19, 2, pp. 112-120, (2014)

[4]

Mehr A.D., Kahya E., Yerdelen C., Linear genetic programming application for successive-station monthly streamflow prediction, Computers and Geosciences, 70, 9, pp. 63-72, (2014)

[5]

Troiano L., Birtolo C., Armenise R., Searching optimal menu layouts by linear genetic programming, Journal of Ambient Intelligence and Humanized Computing, pp. 1-18, (2015)

[6]

Wu C., An improved artificial fish swarm optimization algorithm, CAAI Transactions on Intelligent Systems, 10, 3, pp. 465-469, (2015)

[7]

Brameier M., Banzhaf W., Explicit control of diversity and effective variation distance in linear genetic programming, 5th European Conference on Genetic Programming, pp. 3-5, (2002)

[8]

Gaudesi M., Squillero G., Tonda A., An efficient distance metric for linear genetic programming, 15th Annual Conference on Genetic and Evolutionary Computation, pp. 6-10, (2013)

[9]

Nguyen Q.U., Xuan X.H., O'neill M., Et al., An investigation of fitness sharing with semantic and syntactic distance metrics, Lecture Notes in Computer Science, 7244, pp. 109-120, (2012)

[10]

Tomassini M., Vanneschi L., Collard P., Et al., A study of fitness distance correlation as a difficulty measure in genetic programming, Evolutionary Computation, 13, 2, pp. 213-239, (2005)

← 1 2 →